In light of the reprehensible abuse, including rape threats, received by Caroline Criado-Perez from some Twitter users, there are calls for a “report abuse” button to be added to each tweet. Some have questioned whether this is the right approach on the grounds that it may actually make it easier for those who merely dislike what someone has said to get that person silenced. If Twitter take care to ensure that complaints are properly assessed, including distinguishing between people saying this rashly in the midst of debate and people being genuinely abusive and threatening, then this shouldn’t be a problem. The thing is, implementing such a reporting system will be no mean feat.
To illustrate this, consider that Twitter is generating around 400 million tweets per day (and this figure may grow over time). Assuming that just 1 in 1000 tweets had the “report abuse” button clicked on average, then Twitter would have to handle 400 thousand abuse reports per day, making it a genuine “big data” task. If they were handled by humans (and to minimise the risk of abuse of the feature they should be) and the people employed took on average just 5 mins to process each report, that would entail having 1,389 people employed at all times just to process the tweets. Since humans can’t work 24/7, we’d need to employ at least 3 times the number to ensure 24/7 coverage (assuming 8 hour shifts), and in fact we’d need more than that to take into account holidays, weekends and illness.
Now it may be possible to use automation to reduce the number of reports requiring human handling (if we can identify clear cut cases automatically for example) and to reduce the time spent per tweet, but the more automation there is, the more likely that misuse of the feature may lead to problems or that some cases don’t get properly handled.
That’s not to say it isn’t possible to come up with a system that works (*), just that it’ll require a lot of careful design and testing. But if takes them time to fully implement a new, easier to use system for reporting abuse, this logistical problem may explain why.
(*) Twitter have after all already got a reporting procedure in place, and have started to allow people to report individual tweets from their mobiles.